Challenge: Run an A/B test at work

Not everyone is comfortable testing digital products, but anyone can run design experiments. 

Experimentation matters. It's a good muscle to exercise. Having a clear process for change leads to smarter, less politicized decisions that foster growth in a way that's repeatable and sustainable.

So this week, I thought I'd invite you to run an experiment in the real world. Test a design change in your environment. Here’s how:


A few more things not mentioned in my walk around the block:

  • Obviously, you'll want to somehow be tracking the number of conversions during each testing period. If it's donation boxes you're tracking, for example, empty them out at the time you make the change and be sure to record totals each time you empty them during the 4 weeks.

  • The more significant the change you make, the more likely it is you'll be able to identify real shifts in conversions.

  • Feel free to make more than one change — just don't make changes in the middle of a testing cycle. More variables make it harder to identify the true source of an experiment's success or failure, but for our purposes let's not worry about that much. It's the doing that matters.

  • Sample size would also normally matter — an environment that sees fewer interactions a day will take much longer to reach statistical significance than one that sees hundreds (handy calculator)— but that's not the point of the exercise.

The point is to just dive into testing and see what it feels like to take a measured, principled approach to design/business decisions.

Have a good weekend,

Kyle

Kyle Bowen